InfluxDB 3 Core
InfluxDB 3 Core is an open-source, single-node time series database engine built from the ground up for real-time data workloads.
Publisher review
InfluxDB 3 Core is an open-source, single-node time series database engine built from the ground up for real-time data workloads. Released in April 2025 as a complete rewrite under the FDAP stack (Apache Flight, DataFusion, Arrow, Parquet), it targets developers and operators who need sub-10ms query access to recent telemetry—sensor data, application metrics, or IoT streams—without the overhead of a distributed cluster. Licensed under MIT and Apache 2.0, Core is designed for simplicity: a single binary that starts in seconds and can be upgraded to InfluxDB 3 Enterprise without code changes or data migration.
Core ingests and queries data in milliseconds, delivering sub-10ms access to recent data for real-time dashboards and alerting. It uses columnar compression and object storage (e.g., S3-compatible backends) to cut infrastructure costs on high-volume telemetry, and it supports unlimited cardinality—tracking millions of unique series without performance degradation. A standout capability is running Python scripts directly inside the database engine for in-stream transformations, anomaly detection, and automation, eliminating the need for separate stream processing pipelines. The included Explorer UI provides immediate visualization and querying via SQL.
In the time series database market, InfluxDB 3 Core competes directly with TimescaleDB (PostgreSQL-based), Prometheus (pull-model monitoring), and QuestDB. InfluxData's own benchmarks show Core ingests data approximately 5.5 times slower than QuestDB 8.2.2, though it emphasizes recent-data speed and Python processing over raw throughput. Unlike Prometheus, Core offers full SQL support and unlimited cardinality without sharding. Its separation of compute and storage allows scaling to Enterprise by adding nodes, contrasting with TimescaleDB's single-node PostgreSQL extension model.
The honest trade-offs are significant. InfluxDB 3 Core is an alpha release—not production-ready—and deliberately limited: it only queries data within a maximum 72-hour time span, making it unsuitable for long-term historical analysis. Long-range queries are slower than some competitors (e.g., QuestDB). The single-node design lacks built-in high availability or replication, and the Python processing feature, while innovative, introduces potential security and performance risks in multi-tenant environments. Organizations needing durable storage, compliance, or >72-hour retention must move to the paid Enterprise edition.
How it works
-
Sub-10ms recent data queries
Returns query results in under 10 milliseconds for data within the last 72 hours, enabling real-time dashboards and alerting.
-
Unlimited cardinality support
Tracks millions of unique time series without performance degradation, eliminating the need for series cardinality limits.
-
In-database Python processing
Runs Python scripts directly inside the engine for real-time transformations, anomaly detection, and automation without external pipelines.
-
Columnar compression + object storage
Uses Apache Arrow and Parquet for compression, storing data in S3-compatible object stores to reduce infrastructure costs.
-
FDAP stack architecture
Built on Apache Flight, DataFusion, Arrow, and Parquet for high-performance data transfer, query execution, and columnar storage.
-
SQL and InfluxQL support
Offers full SQL query support alongside InfluxQL, making it accessible to developers familiar with standard relational querying.
-
Single-node simplicity
Runs as a single binary with no cluster management, starting in seconds and upgradable to multi-node Enterprise without rewrites.
Strengths and trade-offs
Strengths
- Open source under MIT and Apache 2.0 licenses, giving developers full freedom to modify and redistribute without restrictions.
- Sub-10ms query latency on recent data, verified by InfluxData benchmarks, making it suitable for real-time monitoring and alerting.
- Compression and object storage reduce storage costs for high-volume telemetry, as columnar Parquet format achieves 10x compression ratios typical for time series.
- Handles unlimited cardinality—millions of unique series—without performance tradeoffs, unlike Prometheus which requires sharding at high cardinality.
Trade-offs
- Alpha release is not production-ready, lacking stability guarantees and full testing for critical deployments.
- Maximum query time span of 72 hours prevents use for long-term historical analysis or compliance-driven data retention.
- Ingest performance is approximately 5.5 times slower than QuestDB 8.2.2, limiting throughput for high-velocity data streams.
- Single-node design provides no built-in high availability or replication, requiring manual failover or migration to Enterprise for redundancy.
Pricing context
Free open-source download for InfluxDB 3 Core. Enterprise edition has custom pricing based on CPU configuration, with a 30-day free trial available. No credit card required for trial.
Getting started with InfluxDB 3 Core
-
Download and run binary
Download the InfluxDB 3 Core binary from the official releases page. Extract the archive and run the executable with default settings. The server starts in seconds and listens on localhost port 8181 for HTTP requests.
-
Connect to object storage
Configure the database to use an S3-compatible object store by setting environment variables for your bucket, region, and access credentials. This stores compressed Parquet data and reduces local storage costs.
-
Define a database and table
Use the InfluxDB 3 CLI or HTTP API to create a database and define a table schema with tags and fields. Specify a measurement name and column types to match your telemetry data structure.
-
Ingest sample telemetry data
Send a batch of time series data using the write API endpoint. Include a timestamp, at least one tag (e.g., sensor_id), and a field (e.g., temperature). Verify ingestion by checking the Explorer UI for new points.
-
Run a SQL query on recent data
Open the Explorer UI or use the query API to execute a SQL SELECT statement filtering by the last hour. Confirm sub-10ms response time for data within the 72-hour window. Adjust filters to test real-time performance.
Frequently Asked Questions
What is InfluxDB 3 Core and what is it used for?
InfluxDB 3 Core is an open-source, single-node time series database for real-time workloads like sensor data, application metrics, and IoT streams. It delivers sub-10ms query access to recent telemetry without cluster overhead, using the FDAP stack.
What are the main features of InfluxDB 3 Core?
Key features include sub-10ms queries on recent data, unlimited cardinality support, in-database Python processing for real-time transformations, columnar compression with object storage, full SQL support, and a single binary that starts in seconds.
How does InfluxDB 3 Core handle unlimited cardinality?
InfluxDB 3 Core tracks millions of unique time series without performance degradation, eliminating the need for cardinality limits. This contrasts with Prometheus, which requires sharding at high cardinality, making Core suitable for diverse telemetry sources.
Can I run Python scripts inside InfluxDB 3 Core?
Yes, InfluxDB 3 Core runs Python scripts directly in the database engine for in-stream transformations, anomaly detection, and automation. This removes the need for separate stream processing pipelines, though it introduces security and performance risks in multi-tenant setups.
What are the limitations of InfluxDB 3 Core as an alpha release?
InfluxDB 3 Core is alpha software, not production-ready, with a maximum query time span of 72 hours, making it unsuitable for long-term historical analysis. It lacks built-in high availability, replication, and ingests data about 5.5 times slower than QuestDB 8.2.2.
How does InfluxDB 3 Core compare to TimescaleDB and Prometheus?
InfluxDB 3 Core offers full SQL support and unlimited cardinality without sharding, unlike Prometheus. It uses a compute-storage separation that scales to Enterprise, contrasting with TimescaleDB's single-node PostgreSQL extension model. Core emphasizes recent-data speed and Python processing over raw throughput.
Alternatives
How InfluxDB 3 Core compares
Direct head-to-head against 3 competitors. Picked by 7wData.
InfluxDB 3 Core
- Pricing
- Free open-source download for InfluxDB 3 Core. Enterprise edition has custom pricing based on CPU configuration, with a 30-day free trial available. No credit card required for trial.
- Target
- InfluxDB 3 Core is an open-source, single-node time series database engine built from the ground up for real-time data workloads.
- Strength
- Open source under MIT and Apache 2.0 licenses, giving developers full freedom to modify and redistribute without restrictions.
- Watch for
- Alpha release is not production-ready, lacking stability guarantees and full testing for critical deployments.
QuestDB
- Pricing
- Open source (Apache 2.0); QuestDB Cloud starts at $0.10/hr
- Target
- High-throughput time-series for finance, IoT, and real-time analytics
- Deployment
- Self-hosted or cloud
- Strength
- Ultra-low latency ingestion and SQL with native Parquet support
- Watch for
- Smaller ecosystem and fewer integrations than InfluxDB
TDengine
- Pricing
- Open source (AGPL); Cloud starts at $0.04/hr
- Target
- IoT, industrial monitoring, and large-scale time-series workloads
- Deployment
- Self-hosted or cloud
- Strength
- Superior query performance over InfluxDB 3 Core in TSBS benchmarks
- Watch for
- InfluxQL compatibility gaps and smaller community than InfluxDB
Prometheus
- Pricing
- Open source (Apache 2.0); no official cloud tier
- Target
- Observability and server metrics with built-in alerting
- Deployment
- Self-hosted only
- Strength
- Dominant CNCF project for Kubernetes-native monitoring
- Watch for
- Limited to pull-based metrics; no SQL support; not for IoT or finance
User reviews
No user reviews yet. Be the first to write one.
Sources
Reporting on this tool draws on these publicly available sources.